Infinite mixture-of-experts model for sparse survival regression with application to breast cancer.

Department of Computer Science, University of Basel, Bernoullistr, 16, CH-4056 Basel, Switzerland.
BMC Bioinformatics (Impact Factor: 2.67). 01/2010; 11 Suppl 8:S8. DOI: 10.1186/1471-2105-11-S8-S8
Source: PubMed

ABSTRACT We present an infinite mixture-of-experts model to find an unknown number of sub-groups within a given patient cohort based on survival analysis. The effect of patient features on survival is modeled using the Cox's proportionality hazards model which yields a non-standard regression component. The model is able to find key explanatory factors (chosen from main effects and higher-order interactions) for each sub-group by enforcing sparsity on the regression coefficients via the Bayesian Group-Lasso.
Simulated examples justify the need of such an elaborate framework for identifying sub-groups along with their key characteristics versus other simpler models. When applied to a breast-cancer dataset consisting of survival times and protein expression levels of patients, it results in identifying two distinct sub-groups with different survival patterns (low-risk and high-risk) along with the respective sets of compound markers.
The unified framework presented here, combining elements of cluster and feature detection for survival analysis, is clearly a powerful tool for analyzing survival patterns within a patient group. The model also demonstrates the feasibility of analyzing complex interactions which can contribute to definition of novel prognostic compound markers.

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This study demonstrates the feasibility of using a modified mixture of experts (ME) model with repeated measured tumoural K(trans) value to perform an automatic diagnosis of responder based on perfusion magnetic resonance imaging (MRI) of rectal cancer. The data used in this study was obtained from 39 patients with primary rectal carcinoma who were scheduled for preoperative chemoradiotherapy. The modified ME model is a joint modeling of the ME model via the linear mixed effect model. First, we considered two local experts and a gating network, and the modified expert network as a liner mixed effect model. Afterward, the finding estimates were obtained via the expectation-maximization algorithm. All computation was performed by R-2.15.2. We found that two experts have different patterns. The feature of expert 1 (n = 10) had a higher baseline value and a lower slope than expert 2 (n = 29). A comparison of the estimated experts and responder/non-responder groups according to T-downstaging criteria showed that expert 1 had a more effect treatment responder than expert 2. A novel feature of this study is that it is an extension of classical ME models in case of repeatedly measured data. The proposed model has the advantages of flexibility and adaptability for identifying distinct subgroups with various time patterns, and it can be applied to biomedical data which is measured repeatedly, such as time-course microarray data or cohort data. This method can assist physicians as important diagnostic decision making mechanism.
    Healthcare informatics research. 06/2013; 19(2):130-6.
  • [Show abstract] [Hide abstract]
    ABSTRACT: The author describes a project in which the goal was to ensure clinician acceptance and use of an expert system that generates information alerts for psychiatric practitioners. The resulting implementation methodology is described, and its potential application in other settings is discussed. The author describes the procedures followed in developing an implementation methodology and key features of the design, and he comments on the applicability of the methodology to other medical settings
    Engineering in Medicine and Biology Society, 1989. Images of the Twenty-First Century., Proceedings of the Annual International Conference of the IEEE Engineering in; 12/1989
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The mixture-of-experts (ME) network uses a modular type of neural network architecture optimized for supervised learning. This model has been applied to a variety of areas related to pattern classification and regression. In this research, we applied a ME model to classify hidden subgroups and test its significance by measuring the stiffness of the liver as associated with the development of liver cirrhosis. The data used in this study was based on transient elastography (Fibroscan) by Kim et al. We enrolled 228 HBsAg-positive patients whose liver stiffness was measured by the Fibroscan system during six months. Statistical analysis was performed by R-2.13.0. A classical logistic regression model together with an expert model was used to describe and classify hidden subgroups. The performance of the proposed model was evaluated in terms of the classification accuracy, and the results confirmed that the proposed ME model has some potential in detecting liver cirrhosis. This method can be used as an important diagnostic decision support mechanism to assist physicians in the diagnosis of liver cirrhosis in patients.
    Healthcare informatics research. 03/2012; 18(1):29-34.

Full-text (2 Sources)

Available from
Jun 1, 2014